基于感知器模型本文提出一种换热器网络优化新算法，在遗传迭代过程中，依据目标函数的适应值，该算法不断调节感知器的权值，使感知器以随机方式搜索网络拓扑结构和运行参数。经过本文的计算表明：这一算法对于多维、多峰、非凸函数也具有很好的适应性，能够有效地克服传统方法（混合整数非线性优化）的不足之处。 Based on the perceptron model, a new algorithm of heat exchanger network optimization is proposed in this paper. During the genetic iteration, according to the fitness of the objective function, the algorithm adjusts the perceptron’s weight continually so that the perceptron searches the network topology in a random manner and Operating parameters. The calculation results show that the proposed algorithm has good adaptability to multidimensional, multimodal and non-convex functions and can effectively overcome the shortcomings of the traditional method (mixed integer nonlinear optimization).